Personal computing has gone, and is still going, through radical changes. Vast increases in computing complexity and faster processing speeds are now available in smaller packages at lower cost. There is, however, a corresponding increase in power consumption to operate modern processor-based devices. This is particularly problematic in portable processor-based devices such as laptop computers and personal data assistants (PDAs) which cannot easily incorporate larger and/or heavier batteries. To extend battery life many such devices implement multiple power conservation states each having a different rate of power consumption, transitioning the device to lower power-consuming states when possible.
Unfortunately, device availability is reduced when the device assumes a state in which limited or no device operations can be performed. In addition, transitioning between certain states requires the performance of tasks that can take considerable time, further decreasing device availability. To avoid this latter loss of device availability, some conventional approaches retain data in processor memory. However, maintaining memory power can unnecessarily consume power, accelerating battery depletion.